To analyse bone stresses in humerus-megaprosthesis construct in response to axial loading under varying implant lengths in proximal humeral replacement following tumour excision.

CT scans of 10 cadaveric humeri were processed in 3D Slicer to obtain three-dimensional (3D) models of the cortical and cancellous bone. Megaprostheses of varying body lengths (L) were modelled in FreeCAD to obtain the 3D geometry. Four FE models: group A consisting of intact bone; groups B (L=40mm), C (L=100mm) and D (L=120mm) comprising of humerus-megaprosthesis constructs were created. Isotropic linear elastic behaviour was assigned for all materials. A tensile load of 200N was applied to the elbow joint surface with the glenohumeral joint fixed with fully bonded contact interfaces. Static analysis was performed in Abaqus. The bone was divided at every 5% bone length beginning distally. Statistical analysis was performed on maximum von Mises stresses in cortical and cancellous bone across each slice using one-way ANOVA (0-45% bone length) and paired t-tests (45-70% bone length). To quantify extent of stress shielding, average percentage change in stress from intact bone was also computed.

Maximum stress was seen to occur distally and anteriorly above the coronoid fossa. Results indicated statistically significant differences between intact state and shorter megaprostheses relative to longer megaprostheses and proximally between intact and implanted bones. Varying levels of stress shielding were recorded across multiple slices for all megaprosthesis lengths. The degree of stress shielding increased with implant lengthening being 2-4 times in C and D compared to B.

Axial loading of the humerus can occur with direct loading on outstretched upper limbs or indirectly through the elbow. Resultant stress shielding effect predicted in longer megaprosthesis models may become clinically relevant in repetitive axial loading during activities of daily living. It is recommended to use shorter megaprosthesis to prevent failure.